MMOCR / mmocr /models /textdet /postprocess /pse_postprocessor.py
tomofi's picture
Add application file
2366e36
raw
history blame
No virus
3.19 kB
# Copyright (c) OpenMMLab. All rights reserved.
import cv2
import numpy as np
import torch
from mmcv.ops import contour_expand
from mmocr.core import points2boundary
from mmocr.models.builder import POSTPROCESSOR
from .base_postprocessor import BasePostprocessor
@POSTPROCESSOR.register_module()
class PSEPostprocessor(BasePostprocessor):
"""Decoding predictions of PSENet to instances. This is partially adapted
from https://github.com/whai362/PSENet.
Args:
text_repr_type (str): The boundary encoding type 'poly' or 'quad'.
min_kernel_confidence (float): The minimal kernel confidence.
min_text_avg_confidence (float): The minimal text average confidence.
min_kernel_area (int): The minimal text kernel area.
min_text_area (int): The minimal text instance region area.
"""
def __init__(self,
text_repr_type='poly',
min_kernel_confidence=0.5,
min_text_avg_confidence=0.85,
min_kernel_area=0,
min_text_area=16,
**kwargs):
super().__init__(text_repr_type)
assert 0 <= min_kernel_confidence <= 1
assert 0 <= min_text_avg_confidence <= 1
assert isinstance(min_kernel_area, int)
assert isinstance(min_text_area, int)
self.min_kernel_confidence = min_kernel_confidence
self.min_text_avg_confidence = min_text_avg_confidence
self.min_kernel_area = min_kernel_area
self.min_text_area = min_text_area
def __call__(self, preds):
"""
Args:
preds (Tensor): Prediction map with shape :math:`(C, H, W)`.
Returns:
list[list[float]]: The instance boundary and its confidence.
"""
assert preds.dim() == 3
preds = torch.sigmoid(preds) # text confidence
score = preds[0, :, :]
masks = preds > self.min_kernel_confidence
text_mask = masks[0, :, :]
kernel_masks = masks[0:, :, :] * text_mask
score = score.data.cpu().numpy().astype(np.float32)
kernel_masks = kernel_masks.data.cpu().numpy().astype(np.uint8)
region_num, labels = cv2.connectedComponents(
kernel_masks[-1], connectivity=4)
labels = contour_expand(kernel_masks, labels, self.min_kernel_area,
region_num)
labels = np.array(labels)
label_num = np.max(labels)
boundaries = []
for i in range(1, label_num + 1):
points = np.array(np.where(labels == i)).transpose((1, 0))[:, ::-1]
area = points.shape[0]
score_instance = np.mean(score[labels == i])
if not self.is_valid_instance(area, score_instance,
self.min_text_area,
self.min_text_avg_confidence):
continue
vertices_confidence = points2boundary(points, self.text_repr_type,
score_instance)
if vertices_confidence is not None:
boundaries.append(vertices_confidence)
return boundaries